A rapid origin identification method for Panax quinquefolius by terahertz spectroscopy combined with random forest

Based on the advantages of terahertz spectroscopy, this paper used terahertz spectroscopy and random forest to identify the origin of Panax quinquefolius. The extinction coefficients ranging from 0.2 to 1.3 THz were applied to establish random forest model. Principal component analysis was used to reduce the dimensionality of the original data and eliminate redundancy and noise. The genetic algorithm was employed to optimize the parameters of the classification model which achieved an identification rate of 90% for the samples. It turns out that terahertz spectroscopy combines with random forest to identify the origin of Panax quinquefolius is rapid, accurate and non-destructive, which is of great significance to the quality control and industrial development of Panax quinquefolius.